This thesis deals with processing of retinal fundus images. Vision is the most important human sense and its injury has very serious consequences for humans. Automatic processing of retinal images increases the efficiency of medical examination and accelerates diagnoses of deseases. Retina exhibits unique characteristics for each person and thus can also be used to identify people. In this task is briefly discussed the structure and properties of each parts of the eye, particularly the retina, and their possible diseases such as diabetic retinopathy, glaucoma and age related macular degeneration. Subsequently, the task describes the representation and characteristics of the digital image. Also is devoted to selected image segmentation methods namely thresholding, edge detection and segmentation techniques based on the matched filter. The outcome of this task is the application in which several segmentation methods are implemented for the blood vessels extraction. For each of these methods it is possible to set the parameters of the segmentation to ensure high quality blood vessels extraction in images of different quality.

This diploma thesis deals with segmentation of blood vessel from images acquired by fundus camera. The characteristic of fundus images and current methods of segmentation are described in theoretical part. The reach of the practical part is method using statistical model. The model using Student´s distribution for automatic segmentation is gradually drafted. Firstly EM- algorithm has been incorporated and model drafted on Markov random fields for improving robustness to noise after that. Contrast of thin blood vessel is improved in image preprocessing part by discrete wave transformation. The output image is used as mask for grayscale intensity decrease of thinnest blood-vessel and intensity increase of background. Whole model was programed in Matlab. The model was tested on whole HRF database. The quantitative evaluation of binary images were compared with golden standard images.

Bakgrund: Diabetesretinopati är en av de vanligaste komplikationerna vid diabetes, och idag finns det en nollvision för allvarlig synskada hos diabetespatienter. Det går att upptäcka diabetesretinopati i ett tidigt skede om regelbunden ögonbottenfotografering görs. Forskningen som finns om ögonbottenfotografering utifrån ögonsjuksköterskans perspektiv är sparsam.  Syfte: var att belysa ögonsjuksköterskans erfarenhet av diabetespatienters besök vid ögonbottenfotografering. Metod: Metoden som användes var en kvalitativ design med ett induktivt tillvägagångssätt. Åtta semistrukturerade intervjuer genomfördes enskilt med ögonsjuksköterskor. En kvalitativ innehållsanalys beskriven av Lundman och Hällgren Graneheim (2017) användes för att analysera data. Resultat: Ögonsjuksköterskorna angav att tiden inte alltid räckte till, att det kunde vara stressigt, samt att många patienter står i kö för ögonbottenfotografering och kön tar inte slut. Ögonsjuksköterskornas erfarenhet var att många patienter inte har kunskap om varför det är viktigt att komma på ögonbottenfotografering och att en del uteblir. Resultatet visade att ögonsjuksköterskorna hade bra kunskap om att bedöma ögonbottenfotograferingar och att ett flertal gav patienten information av resultatet muntligt vid besöket. Resultatet visade även en del positiva erfarenheter, bland annat fördelar med modern teknik samt fördelar i samverkan med andra professioner. Slutsats: Många diabetespatienter har kunskap om diabetes, däremot är det flertal som har okunskap gällande vikten av att gå på regelbunden ögonbottenfotografering. För att upprätthålla nollvisionen för allvarlig synskada hos diabetespatienter behövs mer information till denna patientgrupp. Ögonsjuksköterskan behöver i sin roll som omvårdnadsspecialist mer tid för informationsgivning i mötet med patienten för att ge patienten kunskap om sjukdomen och hur viktigt det är med ögonbottenfotografering. / Background: Diabetic retinopathy is a common diabetes complication, today there is a zero vision for severe visual impairment in diabetic patients. Diabetic retinopathy can be detected early if regular fundus photography is done. Research on fundus photography from the ophthalmic nurse perspective is sparse. Aim: was to illustrate the ophthalmic nurse's experience of diabetes patients' visits for fundus photography. Method: Qualitative design with inductive approach was used where eight semi-structured interviews were conducted individually with ophthalmic nurses. A qualitative content analysis described by Lundman and Hällgren Graneheim (2017) was used for data analysis. Results: The ophthalmic nurses stated that the time was not always enough, it could be stressful, and many patients are queuing to come for fundus photography. The ophthalmic nurses experience was that many patients don´t have knowledge of why it is important to have fundus photography and that some do not attend. Results showed that the ophthalmic nurses had good knowledge of assessing fundus photography and a majority gave the patient information about results verbally during the visit. Results also showed positive experiences, including advantages of modern technology and advantages in cooperation with other professions. Conclusion: Many diabetic patients are knowledgeable about diabetes, but majority are ignorant of the importance of regular fundus photography. To maintain zero vision for severe visual impairment in diabetic patients, more information is needed for these patients. As a nursing specialist, the ophthalmic nurse needs more time with the patient to provide information about the disease and the importance of fundus photography.

Yes / Detecting multiple ocular diseases in fundus images is crucial in ophthalmic diagnosis. This study introduces the Fundus-DeepNet system, an automated multi-label deep learning classification system designed to identify multiple ocular diseases by integrating feature representations from pairs of fundus images (e.g., left and right eyes). The study initiates with a comprehensive image pre-processing procedure, including circular border cropping, image resizing, contrast enhancement, noise removal, and data augmentation. Subsequently, discriminative deep feature representations are extracted using multiple deep learning blocks, namely the High-Resolution Network (HRNet) and Attention Block, which serve as feature descriptors. The SENet Block is then applied to further enhance the quality and robustness of feature representations from a pair of fundus images, ultimately consolidating them into a single feature representation. Finally, a sophisticated classification model, known as a Discriminative Restricted Boltzmann Machine (DRBM), is employed. By incorporating a Softmax layer, this DRBM is adept at generating a probability distribution that specifically identifies eight different ocular diseases. Extensive experiments were conducted on the challenging Ophthalmic Image Analysis-Ocular Disease Intelligent Recognition (OIA-ODIR) dataset, comprising diverse fundus images depicting eight different ocular diseases. The Fundus-DeepNet system demonstrated F1-scores, Kappa scores, AUC, and final scores of 88.56%, 88.92%, 99.76%, and 92.41% in the off-site test set, and 89.13%, 88.98%, 99.86%, and 92.66% in the on-site test set.In summary, the Fundus-DeepNet system exhibits outstanding proficiency in accurately detecting multiple ocular diseases, offering a promising solution for early diagnosis and treatment in ophthalmology. / European Union under the REFRESH – Research Excellence for Region Sustainability and High-tech Industries project number CZ.10.03.01/00/22_003/0000048 via the Operational Program Just Transition. The Ministry of Education, Youth, and Sports of the Czech Republic - Technical University of Ostrava, Czechia under Grants SP2023/039 and SP2023/042.

Yes / Purpose: To demonstrate methods that enable visual field sensitivities to be compared with normative data without restriction to a fixed test pattern. Methods: Healthy participants (n = 60, age 19–50) undertook microperimetry (MAIA-2) using 237 spatially dense locations up to 13° eccentricity. Surfaces were fit to the mean, variance, and 5th percentile sensitivities. Goodness-of-fit was assessed by refitting the surfaces 1000 times to the dataset and comparing estimated and measured sensitivities at 50 randomly excluded locations. A leave-one-out method was used to compare individual data with the 5th percentile surface. We also considered cases with unknown fovea location by adding error sampled from the distribution of relative fovea–optic disc positions to the test locations and comparing shifted data to the fixed surface. Results: Root mean square (RMS) difference between estimated and measured sensitivities were less than 0.5 dB and less than 1.0 dB for the mean and 5th percentile surfaces, respectively. Root mean square differences were greater for the variance surface, median 1.4 dB, range 0.8 to 2.7 dB. Across all participants 3.9% (interquartile range, 1.8–8.9%) of sensitivities fell beneath the 5th percentile surface, close to the expected 5%. Positional error added to the test grid altered the number of locations falling beneath the 5th percentile surface by less than 1.3% in 95% of participants. Conclusions: Spatial interpolation of normative data enables comparison of sensitivity measurements from varied visual field locations. Conventional indices and probability maps familiar from standard automated perimetry can be produced. These methods may enhance the clinical use of microperimetry, especially in cases of nonfoveal fixation.

Yes / Purpose: To investigate how well visual field sensitivity predicts visual acuity at the same locations in macular disease, and to assess whether such predictions may be useful for selecting an optimum area for fixation training. Methods: Visual field sensitivity and acuity were measured at nine locations in the central 10° in 20 people with AMD and stable foveal fixation. A linear mixed model was constructed to predict acuity from sensitivity, taking into account within-subject effects and eccentricity. Cross validation was used to test the ability to predict acuity from sensitivity in a new patient. Simulations tested whether sensitivity can predict nonfoveal regions with greatest acuity in individual patients. Results: Visual field sensitivity (P < 0.0001), eccentricity (P = 0.007), and random effects of subject on eccentricity (P = 0.043) improved the model. For known subjects, 95% of acuity prediction errors (predicted − measured acuity) fell within −0.21 logMAR to +0.18 logMAR (median +0.00 logMAR). For unknown subjects, cross validation gave 95% of acuity prediction errors within −0.35 logMAR to +0.31 logMAR (median −0.01 logMAR). In simulations, the nonfoveal location with greatest predicted acuity had greatest “true” acuity on median 26% of occasions, and median difference in acuity between the location with greatest predicted acuity and the best possible location was +0.14 logMAR (range +0.04 to +0.17). Conclusions: The relationship between sensitivity and acuity in macular disease is not strongly predictive. The location with greatest sensitivity on microperimetry is unlikely to represent the location with the best visual acuity, even if eccentricity is taken into account. / College of Optometrists Postdoctoral Research Award (JD and ATA; London, UK) and National Institute for Health Research (NIHR) Postdoctoral Fellowship (ATA; London, UK). Presents independent research funded by the NIHR. / Research Development Fund Publication Prize Award winner, August 2018.

To use the image of the retina for biometric or medical purposes, it is necessary that the image is of the highest quality and ideally covers the largest possible area. It is therefore necessary to move the sensing device to the most suitable position while ensuring its stability. A laboratory handling platform is used for this purpose. The platform is moved by stepper motors. A control board equipped with a microcontroller has been developed for their control, which allows very fine movements. The theoretical accuracy of the manipulator is up to 20 nm. Since it is necessary for sensing, to have proper retina illumination, the control system also includes adjustable illumination elements in the infrared and visible spectrum. The resulting system allows scanning of the retina of the eye from multiple angles and thereby effectively increase the area on which it is possible to further look for diseases or biometric features.

This work deals with possibilities of registration of retinal images from different mo-dalities, concretely optical coherence tomography (OCT), scanning laser ophthalmoscopy (SLO) and fundus camera. In first stage is the interest focused on registration of SLO and fundus images, which will serve to determine area of interest for consecutive registration of OCT data. The final stage is finding correct location of OCT B-scans in fundus image. On the basis of the studied methods of registration was chosen method making use of computation of correlation coefficient for both cases. For finding optimal parameters of registration is used searching through whole space of parameters. In partial stages of the work was created algorithm for alignment of B-scans followed by detection of blood vessels and also simple algorithm for detection of blood vessels from fundus image. For more transparent registration the graphical user interface was created, which allows loading input images and displaying the result in several possible forms.

Tato práce se zabývá multimodální registrací snímků sítnice z různých skenovacích zařízení. Multimodální registrace umožňuje zvýraznit prvky na snímcích sítnice, které jsou důležité pro detekci různých typů onemocnění oka (jako je glaukom, degradace nervových vláken, degradace cév, atd.). Teoretická část tvoří zhruba první půlku práce a je následována praktickou částí, která popisuje postupy při různých typech registrací snímků z fundus kamery, SLO a OCT. Registrace fundus a SLO snímků je provedena pomocí prostorové transformace. Tato práce popisuje tři různé metody registrace SLO snímků se snímky z fundus kamery. První a zároveň nejjednodušší je manuální registrace. Druhou je automatická registrace založená na metodě korelace. Výsledky, včetně porovnání obou metod, jsou uvedeny v závěru. Třetím typem je poloautomatická registrace, která využívá výhod obou předchozích metod a tím pádem je kompromisem mezi rychlostí a přesností registrace. Registrace fundus snímků a B-scanů z OCT je realizována dvěma různými metodami. První je opět založená na korelaci a druhá na prostorové transformaci. Všechny tyto registrační metody jsou realizovány také prakticky v programovém prostředí Matlab.

The master thesis is focused on creating a methodology for quantification of the nerve fiber layer on photographs of the retina. The introductory part of the text presents a medical motivation of the thesis and mentions several studies dealing with this issue. Furthermore, the work describes available textural features and compares their ability to quantify the thickness of the nerve fiber layer. Based on the described knowledge, the methodology to make different regression models enabling prediction of the retinal nerve fiber layer thickness was developed. Then, the methodology was tested on the available image dataset. The results showed, that the outputs of regression models achieve a high correlation between the predicted output and the retinal nerve fiber layer thickness measured by optical coherence tomography. The conclusion discusses an usability of the applied solution.